maccel works nicely on the Ploopy Mini

[brucala]

Just wanted to mentioned that I tried maccel on a Ploopy Mini (with an ADNS-5050 sensor) and it works beautifully.
I had to tune the parameters, but after I found a good set of parameters maccel made the trackball to finally feel nice on my linux laptop. Before maccel, the the default linux mouse acceleration was totally useless on the Ploopy.

Thanks for maccel.

[burkfers]

I'm glad you like it and that it works so well with the ADNS5050 sensor!

Having to tune the parameters was expected; Our tests concluded that for such a tweakable curve everyone who tried it has different preferences. Did you have to move any parameters very far to get good results on your sensor? I'm interested in whether the parameters even translate at all between different sensor models.
Do you think the documentation should be clearer on the fact that a user is expected to tune the parameters to their liking?

We'd also be happy to have another maccel user over on the home we made for it on the BastardKB discord.

Thank you very much for your feedback!

[brucala]

I think the documentation is great and very clear!
Before finding this maccel implementation I did a bit of research about mouse acceleration in general, and even tried my own implementation in QMK, so I had a head start. But I still think the documentation is fantastic.

I ended up with quite a different parameter set. The default were far for optimal in my case, but I could already tell that this implementation was way smoother and smarter than my rudimentary attempt. This is what I ended up with:

• MACCEL_LIMIT_UPPER 1 -> 2. This is recommended to leave unchanged and instead modify DPI. However, for some reason in my device the max DPI is around 1500. So I ended up with a default DPI of 1000 and a factor 2 here, making an effective max DPI of 2000. Then I can cycle to higher and lower DPIs if i want to.
• MACCEL_LIMIT 0.2 -> 0.12. I found this lower value better for precision at low speed.
• MACCEL_TAKEOFF 2.0. Unchanged
• MACCEL_OFFSET 2.2 -> 1.5. In my device it looks like things are concentrated at low input speeds, so I need acceleration to kick in earlier.
• MACCEL_GROWTH_RATE 0.25 -> 0.65. For the same reason as above, I need to reach the maximum faster.

I wanted to try the via configuration to optimize the parameters but I hit the firmware size limits. So tuned things manually with a trial and error approach. It wasn't difficult because I already had gained some intuition on how the curve should look like when I did my own implementation.